INDEX

Explanations

high-frequency function words indicating relationships and conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

anguard

-0.16

áº£n

-0.14

yas

-0.13

-ra

-0.13

opsis

-0.13

éĩĮ

-0.13

 ones

-0.13

params

-0.12

Tob

-0.12

lus

-0.12

POSITIVE LOGITS

lagen

0.15

0.14

rej

0.14

reesome

0.14

isos

0.14

_ASSUME

0.13

ddb

0.13

oler

0.13

Activations Density 0.391%

high-frequency function words indicating relationships and conditions

No Comments

No Known Activations