INDEX

Explanations

mentions of mistakes or errors

oai_token-act-pair · gpt-3.5-turbo

references to errors or mistakes in various contexts

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

rollers

-0.77

raq

-0.71

electric

-0.71

amen

-0.71

zona

-0.70

well

-0.70

agonists

-0.70

vez

-0.68

azine

-0.67

zie

-0.66

POSITIVE LOGITS

 mistakes

0.87

 mistake

0.76

 careless

0.73

 Gamble

0.72

 forgiven

0.71

fulness

0.70

é»Ĵ

0.70

 perpetrated

0.70

 gamb

0.68

 mishand

0.68

Activations Density 0.051%

mentions of mistakes or errors

references to errors or mistakes in various contexts

No Comments

No Known Activations