INDEX

Explanations

phrases indicating rights or entitlements

oai_token-act-pair · gpt-3.5-turbo

references to individual rights and entitlements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

irie

-0.64

hiba

-0.61

iries

-0.60

 extremes

-0.60

glomer

-0.59

ritz

-0.59

 Madness

-0.57

Byr

-0.57

Ĥª

-0.57

 needless

-0.57

POSITIVE LOGITS

 vested

0.87

ointed

0.81

icum

0.77

 whatsoever

0.76

âĺ

0.70

 veto

0.70

skin

0.69

kees

0.68

atis

0.66

osta

0.64

Activations Density 0.234%

phrases indicating rights or entitlements

references to individual rights and entitlements

No Comments

No Known Activations