INDEX

Explanations

terms related to the design or intended use of various products

oai_token-act-pair · gpt-3.5-turbo

phrases related to design and intended use

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

SPONSORED

-0.75

anon

-0.72

facts

-0.68

most

-0.65

 Thankfully

-0.65

ouf

-0.65

listed

-0.64

rant

-0.63

 deducted

-0.63

natureconservancy

-0.62

POSITIVE LOGITS

geries

0.83

use

0.82

USE

0.82

 purpose

0.82

 optimum

0.79

 consumption

0.72

 optimal

0.71

 maximum

0.71

 sale

0.70

 immortality

0.68

Activations Density 0.159%

terms related to the design or intended use of various products

phrases related to design and intended use

No Comments

No Known Activations