INDEX

Explanations

words related to actions involving force or coercion

oai_token-act-pair · gpt-3.5-turbo

the concept of being restrained or confined

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Karn

-0.75

Nos

-0.70

 agricultural

-0.65

 technologically

-0.65

 analytics

-0.64

 escal

-0.63

GSL

-0.62

Ark

-0.62

 platinum

-0.61

 gateway

-0.61

POSITIVE LOGITS

uff

4.50

uffs

2.57

UFF

2.57

uffle

2.32

uffed

2.15

uf

1.60

uffy

1.55

uffer

1.53

aff

1.18

uck

1.11

Activations Density 0.014%

words related to actions involving force or coercion

the concept of being restrained or confined

No Comments

No Known Activations