INDEX

Explanations

mentions of speeches

oai_token-act-pair · gpt-3.5-turbo

references to speeches

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 Soldiers

-0.65

 Associ

-0.65

rome

-0.62

involved

-0.59

 Mississ

-0.58

 saline

-0.57

 Username

-0.56

 Coast

-0.56

 Users

-0.56

usercontent

-0.56

POSITIVE LOGITS

writers

1.02

writer

1.01

 speeches

0.90

writing

0.80

 speech

0.80

 delivered

0.77

 denouncing

0.77

 imped

0.76

anooga

0.76

speech

0.75

Activations Density 0.043%

mentions of speeches

references to speeches

No Comments

No Known Activations