INDEX

Explanations

terms related to commuting or reputations

oai_token-act-pair · gpt-3.5-turbo

variations of the suffix related to the concepts of communication and reputation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ork

-0.74

luster

-0.65

ware

-0.64

eland

-0.63

aith

-0.61

rompt

-0.60

Ey

-0.60

 Conquest

-0.59

Pers

-0.59

hma

-0.59

POSITIVE LOGITS

anamo

0.92

uted

0.88

uality

0.84

tle

0.84

uting

0.75

arthed

0.74

tarian

0.71

nels

0.71

uated

0.70

iate

0.70

Activations Density 0.013%

terms related to commuting or reputations

variations of the suffix related to the concepts of communication and reputation

No Comments

No Known Activations