INDEX

Explanations

mentions of specific segments within a text or show

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "segment" indicating various parts or divisions of content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 neglig

-0.72

CHA

-0.70

etheless

-0.65

 medically

-0.65

 mechanically

-0.64

gettable

-0.64

 realism

-0.62

kens

-0.62

 mother

-0.62

ccording

-0.62

POSITIVE LOGITS

 segments

1.02

 segment

1.00

regate

0.92

naire

0.89

ation

0.86

eering

0.86

IAL

0.80

arily

0.79

ially

0.78

tainment

0.77

Activations Density 0.005%

mentions of specific segments within a text or show

repeated mentions of the word "segment" indicating various parts or divisions of content

No Comments

No Known Activations