INDEX

Explanations

terms related to uniqueness or distinctiveness

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "unique" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

apers

-0.65

©¶æ

-0.65

vation

-0.63

Rae

-0.63

 PRESS

-0.62

 Alive

-0.61

rollers

-0.60

AFB

-0.60

EEE

-0.59

onest

-0.59

POSITIVE LOGITS

ively

0.99

 identifier

0.91

Magikarp

0.86

 identifiers

0.86

iveness

0.83

 unique

0.82

iates

0.81

isable

0.80

ities

0.80

 uniqueness

0.79

Activations Density 0.021%

terms related to uniqueness or distinctiveness

occurrences of the word "unique" and its variations

No Comments

No Known Activations