INDEX

Explanations

references to popular TV shows and movie franchises

oai_token-act-pair · gpt-3.5-turbo

titles of popular TV shows and movies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

housing

-0.87

UNCH

-0.68

rastructure

-0.65

imentary

-0.65

 DISTR

-0.65

prus

-0.64

onder

-0.64

share

-0.64

UM

-0.64

held

-0.63

POSITIVE LOGITS

Movie

0.97

 sequels

0.95

advertising

0.95

 trilogy

0.94

 Episode

0.89

 Characters

0.88

 trope

0.87

movie

0.87

Synopsis

0.86

 episodes

0.86

Activations Density 0.129%

references to popular TV shows and movie franchises

titles of popular TV shows and movies

No Comments

No Known Activations