INDEX

Explanations

words related to expressing opinions or arguments

oai_token-act-pair · gpt-3.5-turbo

arguments and opinions related to laws and regulations

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

said

-0.73

mentioned

-0.72

 Verse

-0.68

shit

-0.67

eon

-0.66

atoon

-0.66

OTUS

-0.66

Laughs

-0.66

 Announce

-0.65

ukong

-0.65

POSITIVE LOGITS

 unfairly

1.18

 siph

0.91

 underest

0.84

 underestimated

0.83

 overest

0.81

 disproportionately

0.78

 overlooked

0.76

 unfair

0.75

 misunderstood

0.75

 spurred

0.74

Activations Density 0.602%

words related to expressing opinions or arguments

arguments and opinions related to laws and regulations

No Comments

No Known Activations