INDEX

Explanations

mentions of specific locations or organizations

oai_token-act-pair · gpt-3.5-turbo

words and phrases commonly associated with entities or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Petraeus

-0.65

 refreshing

-0.62

 Cald

-0.62

ente

-0.61

 defer

-0.60

 Finder

-0.60

 Logged

-0.59

 underrated

-0.58

 sender

-0.57

ICAN

-0.57

POSITIVE LOGITS

heit

0.92

pta

0.89

å£«

0.80

nikov

0.78

letal

0.77

pport

0.76

efer

0.73

izza

0.71

hett

0.70

hire

0.70

Activations Density 0.209%

mentions of specific locations or organizations

words and phrases commonly associated with entities or organizations

No Comments

No Known Activations