INDEX

Explanations

the letter "t" appearing at the end of words

oai_token-act-pair · gpt-3.5-turbo

negative contractions of the phrase "do not" in various contexts

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

 behavi

-0.76

 Penguin

-0.69

 Radiation

-0.67

Reviewer

-0.66

 Leopard

-0.65

ses

-0.61

 Neigh

-0.60

 disparate

-0.60

 Sharp

-0.59

 Scarlet

-0.59

POSITIVE LOGITS

cha

1.09

ople

0.92

otally

0.89

raq

0.88

unes

0.87

athom

0.86

ude

0.84

âĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢâĶĢ

0.83

 myself

0.81

plet

0.78

Activations Density 0.094%

the letter "t" appearing at the end of words

negative contractions of the phrase "do not" in various contexts

No Comments

No Known Activations