INDEX

Explanations

mentions of specific names, especially "Sandra"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals, particularly those with the name "Sandra"

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ship

-0.94

ships

-0.91

hire

-0.81

alach

-0.78

¥ŀ

-0.73

comings

-0.73

rob

-0.72

quartered

-0.71

emouth

-0.70

irth

-0.69

POSITIVE LOGITS

 Bland

1.02

 Sandra

0.93

 Skinner

0.82

andra

0.82

 Diaz

0.80

 Rivera

0.73

 Lynn

0.71

 Flores

0.71

 Gomez

0.71

 Barnett

0.70

Activations Density 0.011%

mentions of specific names, especially "Sandra"

references to specific individuals, particularly those with the name "Sandra"

No Comments

No Known Activations