INDEX

Explanations

adjectives and phrases related to underlying issues or problems

oai_token-act-pair · gpt-3.5-turbo

references to fundamental issues or causes behind various subjects

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

ooters

-0.86

asia

-0.84

zzi

-0.80

nery

-0.75

itto

-0.74

kie

-0.74

cker

-0.73

efully

-0.73

cture

-0.72

oping

-0.71

POSITIVE LOGITS

 underlying

0.85

 underpin

0.82

 assumptions

0.80

 motivations

0.79

SourceFile

0.78

 fundamentals

0.77

 infrastructure

0.76

Topic

0.71

 layers

0.70

 structures

0.69

Activations Density 0.015%

adjectives and phrases related to underlying issues or problems

references to fundamental issues or causes behind various subjects

No Comments

No Known Activations