INDEX

Explanations

instances where daring or challenging actions are described

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "dare" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

urgy

-0.79

ensive

-0.75

effic

-0.75

ĻĤ

-0.70

division

-0.69

ounced

-0.64

 Methods

-0.62

 anesthesia

-0.62

edged

-0.62

Tool

-0.61

POSITIVE LOGITS

 Dare

1.35

 dare

1.20

 daring

0.77

 dared

0.76

 boldly

0.74

 provoke

0.72

 censor

0.70

cot

0.68

geon

0.67

 defy

0.67

Activations Density 0.008%

instances where daring or challenging actions are described

occurrences of the word "dare" in various contexts

No Comments

No Known Activations