INDEX

Explanations

instructions or descriptions related to covering something

oai_token-act-pair · gpt-3.5-turbo

terms related to covering or protection

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

speaking

-0.73

ZI

-0.72

 Rebels

-0.72

ĸ

-0.68

 Sabha

-0.67

cgi

-0.63

 reincarn

-0.63

 resurg

-0.62

 rhet

-0.61

erers

-0.61

POSITIVE LOGITS

alls

0.92

 covering

0.86

ages

0.83

 layer

0.82

 blanket

0.82

 cover

0.81

 covers

0.77

 insulation

0.77

 layers

0.76

 blankets

0.71

Activations Density 0.050%

instructions or descriptions related to covering something

terms related to covering or protection

No Comments

No Known Activations