INDEX

Explanations

references to submarines

oai_token-act-pair · gpt-3.5-turbo

references to submarines and submarine-related terms

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

 Marketable

-0.78

âķĲâķĲ

-0.77

place

-0.76

giving

-0.74

ellen

-0.73

phe

-0.73

eful

-0.72

kee

-0.71

td

-0.70

cb

-0.70

POSITIVE LOGITS

 submarine

1.21

 submarines

1.20

 submar

1.03

marine

0.93

 submer

0.86

 torped

0.85

 penetration

0.84

 submerged

0.84

iltration

0.82

 bomber

0.82

Activations Density 0.005%

references to submarines

references to submarines and submarine-related terms

No Comments

No Known Activations