INDEX

Explanations

phrases introducing lists or examples

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the presence of statistics or information being presented

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

rouse

-0.66

culosis

-0.65

iosity

-0.65

odder

-0.63

sense

-0.63

matter

-0.62

amphetamine

-0.61

sed

-0.61

Brexit

-0.61

iet

-0.60

POSITIVE LOGITS

 screenshots

0.81

 excerpt

0.72

 examples

0.71

 snapshots

0.68

 tentative

0.66

nces

0.65

NOT

0.65

 Shooter

0.64

 hoping

0.63

wolves

0.62

Activations Density 0.027%

phrases introducing lists or examples

phrases indicating the presence of statistics or information being presented

No Comments

No Known Activations