INDEX

Explanations

phrases related to political leadership and decision-making

oai_token-act-pair · gpt-3.5-turbo

references to governmental or political structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 namely

-0.79

 ventured

-0.75

 suppose

-0.75

 wondered

-0.75

 seemed

-0.74

 differed

-0.74

 existed

-0.70

 tended

-0.69

 wore

-0.69

 consisted

-0.68

POSITIVE LOGITS

 next

1.41

 remainder

1.37

 upcoming

1.29

1.18

 forthcoming

1.15

 foreseeable

1.14

 entirety

1.11

 future

1.09

1.04

Activations Density 0.596%

phrases related to political leadership and decision-making

references to governmental or political structures

No Comments

No Known Activations