INDEX

Explanations

mentions of locations/cities

oai_token-act-pair · gpt-3.5-turbo

references to political events or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Escape

-0.81

chwitz

-0.81

LW

-0.78

 Shades

-0.78

asa

-0.77

edience

-0.75

sq

-0.72

 Swanson

-0.71

aughs

-0.70

ullivan

-0.69

POSITIVE LOGITS

awei

0.68

 bragging

0.68

 reputation

0.65

 disadvant

0.64

ver

0.63

 bottleneck

0.60

 plag

0.59

 princip

0.59

KNOWN

0.59

 disl

0.58

Activations Density 0.000%

mentions of locations/cities

references to political events or situations

No Comments

No Known Activations