INDEX

Explanations

people's professions or political affiliations

oai_token-act-pair · gpt-3.5-turbo

terms and titles associated with political affiliations and professions

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

assies

-0.71

ravings

-0.69

 timelines

-0.68

Tokens

-0.68

 Dalai

-0.66

 Parties

-0.64

 sockets

-0.62

zees

-0.61

 VIDEOS

-0.60

gements

-0.60

POSITIVE LOGITS

 planner

0.87

 representative

0.74

 adjunct

0.74

 specializing

0.71

hunter

0.71

 affiliate

0.70

 assistant

0.69

querque

0.68

 attorney

0.67

 enthusiast

0.67

Activations Density 0.373%

people's professions or political affiliations

terms and titles associated with political affiliations and professions

No Comments

No Known Activations