INDEX

Explanations

mentions of individuals holding positions of authority or leadership

oai_token-act-pair · gpt-3.5-turbo

references to leadership positions, specifically chairs of committees or organizations

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

 IMAGES

-0.91

 condu

-0.63

 DATA

-0.62

ogue

-0.61

DEF

-0.59

 rugged

-0.59

 JUSTICE

-0.59

ILLE

-0.59

 conver

-0.59

 suscept

-0.59

POSITIVE LOGITS

chair

1.19

person

1.13

woman

1.13

 chair

1.11

manship

1.08

mans

1.05

lift

1.04

chairs

1.02

holder

1.01

pins

0.97

Activations Density 0.016%

mentions of individuals holding positions of authority or leadership

references to leadership positions, specifically chairs of committees or organizations

No Comments

No Known Activations