INDEX

Explanations

mentions of specific names, particularly related to political figures

oai_token-act-pair · gpt-3.5-turbo

references to Donald Trump and related political activity

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

rica

-0.74

eret

-0.74

hap

-0.69

ilver

-0.67

andre

-0.65

laus

-0.64

ruary

-0.63

arch

-0.62

arts

-0.61

lag

-0.61

POSITIVE LOGITS

 Recomm

0.73

Construct

0.60

 holiest

0.59

 explodes

0.59

Senate

0.57

 tariffs

0.57

 reacts

0.57

 interference

0.56

 quitting

0.54

 indu

0.54

Activations Density 0.036%

mentions of specific names, particularly related to political figures

references to Donald Trump and related political activity

No Comments

No Known Activations