INDEX

Explanations

terms related to government officials and political positions

oai_token-act-pair · gpt-3.5-turbo

terms related to government officials and political titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ynski

-0.88

âĸĪâĸĪâĸĪâĸĪ

-0.79

lando

-0.78

erness

-0.77

uckle

-0.75

ylan

-0.74

CRIP

-0.73

ibel

-0.72

 Becky

-0.70

âĸĪâĸĪ

-0.70

POSITIVE LOGITS

 Gujarat

1.15

BJP

1.13

 Delhi

1.12

 Maharashtra

1.11

 Uttar

1.05

 Bihar

1.04

 Bengal

1.01

 Modi

0.99

jriwal

0.97

 Pradesh

0.97

Activations Density 0.208%

terms related to government officials and political positions

terms related to government officials and political titles

No Comments

No Known Activations