INDEX

Explanations

names related to people

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

PDATE

-0.85

vernment

-0.83

ptives

-0.79

culosis

-0.73

GROUND

-0.72

lder

-0.69

lda

-0.67

FUL

-0.66

 firearms

-0.64

pmwiki

-0.63

POSITIVE LOGITS

arine

1.02

ean

0.89

aniel

0.86

ie

0.85

opher

0.83

leen

0.82

ees

0.81

wyn

0.81

imer

0.81

len

0.79

Activations Density 0.014%

names related to people

proper nouns, specifically names of individuals

No Comments

No Known Activations