INDEX

Explanations

phrases related to work and effort in different contexts

oai_token-act-pair · gpt-3.5-turbo

discussions about game development and industry dynamics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 vomiting

-0.57

 undet

-0.57

obar

-0.54

 chlorine

-0.54

Fax

-0.54

azo

-0.54

 voic

-0.54

 landsc

-0.53

 unlaw

-0.52

awi

-0.52

POSITIVE LOGITS

genre

0.68

 Indie

0.66

 Nerd

0.62

omics

0.60

 Gleaming

0.60

lisher

0.59

 academia

0.59

 Freak

0.59

ivals

0.58

leases

0.58

Activations Density 1.811%

phrases related to work and effort in different contexts

discussions about game development and industry dynamics

No Comments

No Known Activations