INDEX

Explanations

instances where something is being checked or examined

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "checked."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

thora

-0.69

iens

-0.66

ennes

-0.64

 parallels

-0.62

nery

-0.62

olution

-0.59

iasm

-0.59

DEM

-0.58

reement

-0.58

 Lever

-0.57

POSITIVE LOGITS

 checked

3.63

checked

2.61

 checking

1.87

 check

1.83

 checks

1.81

 inspected

1.80

checking

1.66

check

1.60

checks

1.56

 tested

1.55

Activations Density 0.009%

instances where something is being checked or examined

repeated mentions of the word "checked."

No Comments

No Known Activations