INDEX

Explanations

mentions of interactions and conversations involving various topics or individuals

oai_token-act-pair · gpt-3.5-turbo

repetitive phrases or conjunctions that indicate a continuation of thoughts or ideas

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥĥãĥī

-0.79

ãĥģ

-0.71

erenn

-0.71

mite

-0.68

ummer

-0.67

ãĥŁ

-0.66

orm

-0.65

atten

-0.65

ãĥł

-0.65

Tier

-0.63

POSITIVE LOGITS

how

1.77

why

1.52

how

1.33

 whether

1.30

why

1.22

 what

1.18

 wondered

1.07

WHY

1.01

HOW

1.00

 whence

0.96

Activations Density 0.284%

mentions of interactions and conversations involving various topics or individuals

repetitive phrases or conjunctions that indicate a continuation of thoughts or ideas

No Comments

No Known Activations