INDEX

Explanations

references indicated by the word "See"

oai_token-act-pair · gpt-3.5-turbo

references to other sources or citations

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

cffffcc

-0.88

EStreamFrame

-0.79

ascal

-0.74

intendent

-0.69

operated

-0.69

aced

-0.67

icum

-0.67

ivari

-0.66

purpose

-0.65

 whistle

-0.64

POSITIVE LOGITS

 supra

0.85

KER

0.76

 Also

0.75

urat

0.71

 below

0.70

 ALSO

0.69

hen

0.69

 Sick

0.68

hee

0.68

lights

0.68

Activations Density 0.045%

references indicated by the word "See"

references to other sources or citations

No Comments

No Known Activations