INDEX

Explanations

names of people and organizations

oai_token-act-pair · gpt-3.5-turbo

mentions of notable individuals or names related to entertainment or media

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

Tokens

-0.70

Xi

-0.69

 respectively

-0.68

depending

-0.65

cause

-0.63

Flight

-0.62

Plex

-0.60

Compare

-0.60

Fed

-0.60

ADRA

-0.60

POSITIVE LOGITS

 writes

0.91

 Photography

0.91

 Profile

0.89

 Quote

0.87

Jr

0.87

 Interview

0.86

(@

0.85

 wrote

0.85

 discusses

0.81

aka

0.79

Activations Density 0.293%

names of people and organizations

mentions of notable individuals or names related to entertainment or media

No Comments

No Known Activations