INDEX

Explanations

words related to specific names and terms, likely categorized by their unique spellings

oai_token-act-pair · gpt-3.5-turbo

abbreviations or acronyms related to topics or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 sympathetic

-0.73

 compensated

-0.72

 unfavorable

-0.70

 privately

-0.69

 awarding

-0.69

 favorable

-0.68

 fortunate

-0.65

 nationally

-0.65

 encouraging

-0.65

 credential

-0.65

POSITIVE LOGITS

phia

1.37

idae

1.29

tein

1.24

opsis

1.23

aceae

1.07

aten

1.06

hea

1.06

ania

1.05

tics

1.04

cs

1.03

Activations Density 0.285%

words related to specific names and terms, likely categorized by their unique spellings

abbreviations or acronyms related to topics or entities

No Comments

No Known Activations