INDEX

Explanations

terms related to warriors or fighting

oai_token-act-pair · gpt-3.5-turbo

references to "warriors," indicating a focus on themes of strength, combat, or heroic figures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

uate

-0.77

ĸļ

-0.76

orage

-0.75

ories

-0.72

alez

-0.70

elong

-0.70

election

-0.67

upon

-0.65

perm

-0.65

sembly

-0.65

POSITIVE LOGITS

riors

1.29

rior

1.20

 warriors

0.98

 warrior

0.94

llan

0.90

Â¯Â¯Â¯Â¯

0.84

fare

0.80

fish

0.76

jit

0.75

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

0.74

Activations Density 0.013%

terms related to warriors or fighting

references to "warriors," indicating a focus on themes of strength, combat, or heroic figures

No Comments

No Known Activations