INDEX

Explanations

contractions

oai_token-act-pair · gpt-3.5-turbo

negations, references to past actions, and statements of belief or assertion

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

emetery

-0.77

 millenn

-0.72

OTAL

-0.70

acebook

-0.66

illion

-0.65

ategory

-0.65

animal

-0.63

ugal

-0.63

icultural

-0.63

illions

-0.63

POSITIVE LOGITS

 Tsarnaev

0.69

Sud

0.64

 Cube

0.63

henko

0.61

rams

0.60

 characterization

0.60

 Chak

0.60

arnaev

0.58

 herself

0.58

 Wasserman

0.58

Activations Density 0.800%

contractions

negations, references to past actions, and statements of belief or assertion

No Comments

No Known Activations