INDEX

Explanations

phrases that denote attributes, qualities, and characteristics of objects or concepts

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

The neuron activates for the word "that" when it introduces a relative clause that describes or provides additional information about a preceding noun or concept.

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

subordinate clause markers, especially the complementizer introducing relative or content clauses.

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

eya

-0.17

Å¡ak

-0.16

/library

-0.15

ajs

-0.14

-License

-0.14

 Holt

-0.14

empo

-0.14

Î

-0.14

eyn

-0.14

okit

-0.14

POSITIVE LOGITS

appropri

0.16

few

0.15

 appropri

0.15

taire

0.15

enco

0.14

 Birmingham

0.14

imon

0.14

 Reactive

0.14

↵

0.13

Activations Density 0.120%

phrases that denote attributes, qualities, and characteristics of objects or concepts

The neuron activates for the word "that" when it introduces a relative clause that describes or provides additional information about a preceding noun or concept.

subordinate clause markers, especially the complementizer introducing relative or content clauses.

No Comments

No Known Activations