INDEX

Explanations

the word "comp" or partial matches such as "complicated" or "company."

oai_token-act-pair · gpt-3.5-turbo

words associated with the term "Gomp"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

piece

-0.70

Interstitial

-0.69

kar

-0.68

cius

-0.67

 Sabha

-0.66

nery

-0.64

 Rite

-0.63

tty

-0.63

 void

-0.62

warming

-0.61

POSITIVE LOGITS

omp

1.31

onent

1.19

odcast

1.10

ortun

1.04

rehens

0.99

oint

0.94

ression

0.93

atible

0.92

ressor

0.90

ensable

0.89

Activations Density 0.005%

the word "comp" or partial matches such as "complicated" or "company."

words associated with the term "Gomp"

No Comments

No Known Activations