INDEX

Explanations

descriptions of specific nations

oai_token-act-pair · gpt-3.5-turbo

references to nations and countries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Accessory

-0.92

elight

-0.76

odiac

-0.75

uce

-0.72

EngineDebug

-0.71

ĸļ

-0.71

idth

-0.69

 showc

-0.66

 guiActiveUn

-0.66

DERR

-0.65

POSITIVE LOGITS

builder

0.85

hood

0.74

 governments

0.72

abad

0.71

wide

0.71

 owned

0.70

 founded

0.69

 bombed

0.68

 headquartered

0.68

 multiplier

0.68

Activations Density 0.243%

descriptions of specific nations

references to nations and countries

No Comments

No Known Activations