INDEX

Explanations

technical terms and proper names likely related to a specific topic or field

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

STD

-0.81

 referen

-0.77

 exha

-0.75

 challeng

-0.73

 reactive

-0.73

 carbohyd

-0.72

uckland

-0.70

 glim

-0.70

NPR

-0.69

bern

-0.67

POSITIVE LOGITS

tein

0.96

 Parish

0.93

hire

0.93

Â®

0.90

 Lines

0.87

 Pact

0.85

 Plaza

0.85

 Â®

0.82

 Palace

0.80

Â®,

0.78

Activations Density 0.319%

technical terms and proper names likely related to a specific topic or field

No Comments

No Known Activations