INDEX

Explanations

mentions that something is widely known, believed, or speculated about

oai_token-act-pair · gpt-3.5-turbo

phrases related to being widely recognized or reported

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

ourke

-0.82

cair

-0.79

 Bagg

-0.72

 Odyssey

-0.70

fter

-0.68

 Gemini

-0.67

ilan

-0.67

ÄŁ

-0.66

 Keeper

-0.66

LV

-0.65

POSITIVE LOGITS

 spaced

0.95

 dispersed

0.91

spread

0.90

 regarded

0.90

 distributed

0.89

 publicized

0.84

 spread

0.82

 dissemin

0.81

entimes

0.79

soType

0.76

Activations Density 0.016%

mentions that something is widely known, believed, or speculated about

phrases related to being widely recognized or reported

No Comments

No Known Activations