INDEX

Explanations

words relating to actions or instructions, possibly with associated verbs or objects

oai_token-act-pair · gpt-3.5-turbo

action verbs related to various tasks or activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CRE

-0.71

lit

-0.62

SPONSORED

-0.62

terday

-0.61

 disg

-0.58

PHOTOS

-0.56

udeb

-0.56

artney

-0.55

prototype

-0.54

tnc

-0.54

POSITIVE LOGITS

 yourselves

1.41

 yourself

1.35

 Yourself

1.15

 your

1.14

ing

1.04

able

0.93

ables

0.92

ments

0.91

ify

0.91

ings

0.91

Activations Density 0.354%

words relating to actions or instructions, possibly with associated verbs or objects

action verbs related to various tasks or activities

No Comments

No Known Activations