INDEX

Explanations

phrases related to being directed towards a target or purpose

oai_token-act-pair · gpt-3.5-turbo

instances of the word "directed."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etus

-0.77

TON

-0.74

board

-0.72

-0.71

 Coch

-0.70

 Organization

-0.70

urity

-0.70

Face

-0.69

Du

-0.68

tex

-0.68

POSITIVE LOGITS

 directed

1.18

 directing

0.99

irect

0.97

 directs

0.92

irection

0.89

 redirect

0.81

ensibly

0.78

 cinem

0.78

 theat

0.77

 cinema

0.75

Activations Density 0.008%

phrases related to being directed towards a target or purpose

instances of the word "directed."

No Comments

No Known Activations