INDEX

Explanations

the word "Bing"

oai_token-act-pair · gpt-3.5-turbo

mentions of the search engine Bing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

terday

-0.84

CLASSIFIED

-0.76

aunder

-0.72

å£«

-0.71

uador

-0.68

ktop

-0.68

OPE

-0.68

ULT

-0.66

ãĤ®

-0.65

escal

-0.64

POSITIVE LOGITS

 Bing

1.32

ham

1.09

 Crosby

1.02

aret

0.86

 Ning

0.85

bles

0.84

tons

0.83

enstein

0.82

bing

0.82

legate

0.80

Activations Density 0.007%

the word "Bing"

mentions of the search engine Bing

No Comments

No Known Activations