INDEX

Explanations

exceptions to a general rule or trend

oai_token-act-pair · gpt-3.5-turbo

phrases indicating exceptions or exclusions

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

 neighb

-0.71

 rall

-0.67

 cart

-0.65

 ancest

-0.63

oded

-0.61

eny

-0.61

ger

-0.61

 precinct

-0.59

 parcel

-0.58

eng

-0.58

POSITIVE LOGITS

 exceptions

0.78

dylib

0.76

DragonMagazine

0.70

blance

0.70

 Legendary

0.69

rities

0.69

ptions

0.66

ĨĴ

0.66

ngth

0.65

perty

0.65

Activations Density 0.015%

exceptions to a general rule or trend

phrases indicating exceptions or exclusions

No Comments

No Known Activations