INDEX

Explanations

terms related to genres in entertainment such as movies, fiction, and games

oai_token-act-pair · gpt-3.5-turbo

references to different genres in entertainment

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

urion

-0.93

 Lumpur

-0.80

loo

-0.79

vic

-0.75

riel

-0.75

administ

-0.72

erald

-0.71

amen

-0.70

ilon

-0.69

Fac

-0.67

POSITIVE LOGITS

 fiction

0.92

¥µ

0.83

 genres

0.83

 tropes

0.83

 mash

0.81

 conventions

0.81

ologies

0.81

genre

0.81

icity

0.78

 genre

0.77

Activations Density 0.040%

terms related to genres in entertainment such as movies, fiction, and games

references to different genres in entertainment

No Comments

No Known Activations