INDEX

Explanations

mentions of political positions or government roles

oai_token-act-pair · gpt-3.5-turbo

references to political titles and positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

utic

-0.71

zsche

-0.68

 Kills

-0.67

 Shades

-0.66

 cipher

-0.65

 Cock

-0.64

anium

-0.63

 ROCK

-0.62

 curl

-0.61

 Drunk

-0.61

POSITIVE LOGITS

president

0.79

 Olymp

0.72

exec

0.71

Prime

0.70

Registered

0.70

boss

0.69

legal

0.67

develop

0.67

joining

0.67

Else

0.67

Activations Density 0.058%

mentions of political positions or government roles

references to political titles and positions

No Comments

No Known Activations