INDEX

Explanations

mentions of the word "fox."

oai_token-act-pair · gpt-3.5-turbo

references to foxes and terms related to minimization and maximization

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

Rated

-0.82

Interstitial

-0.73

Effective

-0.70

INK

-0.69

 ISIL

-0.67

CI

-0.66

 immersion

-0.65

 Mour

-0.64

 Franch

-0.64

externalActionCode

-0.64

POSITIVE LOGITS

ours

0.94

bour

0.83

clud

0.81

es

0.80

holes

0.76

pex

0.76

yip

0.75

erb

0.75

ples

0.74

elled

0.74

Activations Density 0.033%

mentions of the word "fox."

references to foxes and terms related to minimization and maximization

No Comments

No Known Activations