INDEX

Explanations

dates in textual format

oai_token-act-pair · gpt-3.5-turbo

the end of a text or document

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 EVERY

-0.88

 goddamn

-0.86

 darn

-0.83

 damn

-0.82

 damned

-0.77

 fucking

-0.76

#$

-0.75

 infinitely

-0.75

 dudes

-0.74

 ALWAYS

-0.73

POSITIVE LOGITS

resa

1.04

nsic

0.93

iday

0.90

odore

0.89

stanbul

0.86

respond

0.84

 spokeswoman

0.83

intendent

0.80

anmar

0.79

swers

0.79

Activations Density 0.356%

dates in textual format

the end of a text or document

No Comments

No Known Activations