INDEX

Explanations

terms related to emotional states and their implications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

Looking at the activations, this neuron activates on tokens that appear to be at word/phrase boundaries or transitions, particularly where there are structural or grammatical shifts in the text. The strongest activations (0.449, 0.436, 0.422, 0

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

common function words and generic abstract nouns, weakly signaling neutral or expository language.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 15-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L15R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.15.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

odel

-0.17

ÅĻÃŃt

-0.15

dob

-0.15

ÑıÐ±

-0.14

elage

-0.14

é¦Ļèķī

-0.14

enberg

-0.14

ORM

-0.14

ãİ

-0.14

eler

-0.14

POSITIVE LOGITS

etur

0.19

vur

0.17

 Invent

0.16

ixo

0.15

Det

0.15

adin

0.14

VENTORY

0.14

oid

0.14

 sire

0.13

endet

0.13

Activations Density 0.001%

terms related to emotional states and their implications

Looking at the activations, this neuron activates on tokens that appear to be at word/phrase boundaries or transitions, particularly where there are structural or grammatical shifts in the text. The strongest activations (0.449, 0.436, 0.422, 0

common function words and generic abstract nouns, weakly signaling neutral or expository language.

No Comments

No Known Activations