INDEX

Explanations

specific mentions of the word 'fact'

oai_token-act-pair · gpt-3.5-turbo

the word "fact" and its variations, indicating a focus on statements of truth or assertion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

neys

-0.65

 unlocks

-0.62

yles

-0.61

ike

-0.60

ware

-0.59

skin

-0.59

Default

-0.58

otor

-0.58

fred

-0.58

meal

-0.58

POSITIVE LOGITS

 fact

3.77

fact

2.09

 Fact

1.93

Fact

1.85

 facts

1.67

 truth

1.55

 reality

1.43

facts

1.32

 Facts

1.28

 irony

1.18

Activations Density 0.034%

specific mentions of the word 'fact'

the word "fact" and its variations, indicating a focus on statements of truth or assertion

No Comments

No Known Activations