INDEX

Explanations

descriptions related to physical surroundings and objects

oai_token-act-pair · gpt-3.5-turbo

detailing and imagery related to scenes and objects

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

busters

-0.86

olition

-0.74

 boycot

-0.74

 taxpayer

-0.73

ãĤ´ãĥ³

-0.72

ilater

-0.69

 Nationwide

-0.69

 bandwagon

-0.68

 stadiums

-0.67

 airports

-0.67

POSITIVE LOGITS

 protr

1.13

 faint

1.06

 dangling

0.99

 indist

0.96

 faintly

0.94

 pierced

0.93

 slit

0.92

 translucent

0.91

 beck

0.90

 inscribed

0.89

Activations Density 0.480%

descriptions related to physical surroundings and objects

detailing and imagery related to scenes and objects

No Comments

No Known Activations