INDEX

Explanations

actions or instructions directed at the reader

oai_token-act-pair · gpt-3.5-turbo

commands and actions directed towards the reader

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

HCR

-0.75

burgh

-0.63

DragonMagazine

-0.62

EStreamFrame

-0.62

 await

-0.60

Sys

-0.59

OX

-0.58

esi

-0.58

 awaited

-0.58

enic

-0.57

POSITIVE LOGITS

 yourself

0.92

 correctly

0.77

 enough

0.77

 somebody

0.77

 yourselves

0.75

 incorrectly

0.70

 your

0.69

 google

0.67

 anymore

0.66

 someone

0.66

Activations Density 0.208%

actions or instructions directed at the reader

commands and actions directed towards the reader

No Comments

No Known Activations