INDEX

Explanations

movie titles

oai_token-act-pair · gpt-3.5-turbo

references to superhero movies and related franchises

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

heet

-0.76

iery

-0.70

kered

-0.66

 paddle

-0.63

anguage

-0.63

GY

-0.62

ullivan

-0.61

hap

-0.61

imposed

-0.60

bed

-0.60

POSITIVE LOGITS

Of

1.04

 Trilogy

1.03

 Crusade

0.92

 Wars

0.89

 Heist

0.89

 Awakens

0.88

 Hunters

0.87

 Origins

0.87

 Resurrection

0.84

 Returns

0.84

Activations Density 0.164%

movie titles

references to superhero movies and related franchises

No Comments

No Known Activations