INDEX

Explanations

phrases emphasizing arguments and analyses, particularly focusing on their validity and potential flaws

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

New Auto-Interp

Configuration

Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post

Features

65,536

Data Type

float32

Hook Name

blocks.25.hook_resid_post

Hook Layer

Architecture

gated

Context Size

1,024

Dataset

Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alian

-0.15

uspended

-0.15

mate

-0.14

ilet

-0.14

kus

-0.14

 liar

-0.14

ergus

-0.14

onas

-0.14

-regexp

-0.14

oeff

-0.13

POSITIVE LOGITS

 overlook

0.45

 forget

0.40

 forgetting

0.36

 neglect

0.36

 overlooking

0.36

å¿ĺ

0.36

 ignore

0.35

 ignores

0.34

forget

0.34

 misses

0.33

Activations Density 0.110%

phrases emphasizing arguments and analyses, particularly focusing on their validity and potential flaws

No Comments

No Known Activations