INDEX

Explanations

instances of words related to colors

oai_token-act-pair · gpt-3.5-turbo

references to racial and cultural identifiers or stereotypes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

igious

-0.72

FUN

-0.67

Services

-0.66

izarre

-0.66

atis

-0.65

Internal

-0.64

Effective

-0.64

Specific

-0.64

bably

-0.64

ãĤ¡

-0.63

POSITIVE LOGITS

 stripes

0.92

 striped

0.80

 stripe

0.79

oxide

0.77

 flakes

0.76

 Metallic

0.74

 syndrome

0.73

cloth

0.73

 stain

0.71

berries

0.70

Activations Density 0.303%

instances of words related to colors

references to racial and cultural identifiers or stereotypes

No Comments

No Known Activations