INDEX

Explanations

references to predecessors, particularly in political or professional contexts

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "predecessor" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arov

-0.85

oros

-0.82

Jew

-0.69

Ku

-0.67

ux

-0.67

Aid

-0.66

Money

-0.66

ering

-0.65

kell

-0.64

erer

-0.64

POSITIVE LOGITS

 predecessors

0.94

 predecessor

0.92

deen

0.78

Tad

0.77

 generations

0.76

cules

0.74

 successor

0.68

hood

0.68

thood

0.68

etheless

0.66

Activations Density 0.017%

references to predecessors, particularly in political or professional contexts

mentions of the word "predecessor" and related terms

No Comments

No Known Activations