INDEX

Explanations

phrases related to the speaker's own actions and behaviors

oai_token-act-pair · gpt-3.5-turbo

phrases related to the speaker's state of being and presence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

pedia

-0.77

izable

-0.75

sonian

-0.68

=-=-=-=-

-0.65

IZE

-0.65

\\\\\\\\

-0.64

osate

-0.63

 rarity

-0.61

kamp

-0.60

atto

-0.60

POSITIVE LOGITS

 confronted

1.05

 tempted

1.04

 done

1.03

 asleep

1.03

 finished

0.97

 ready

0.96

 bored

0.94

 asked

0.93

 able

0.93

 bitten

0.90

Activations Density 0.117%

phrases related to the speaker's own actions and behaviors

phrases related to the speaker's state of being and presence

No Comments

No Known Activations