INDEX

Explanations

descriptive and impactful words related to various situations and objects

oai_token-act-pair · gpt-3.5-turbo

terms related to negative or challenging situations

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

ernels

-0.64

events

-0.63

 bells

-0.57

rooms

-0.57

 Cups

-0.57

 interfaces

-0.56

 channels

-0.56

 embassies

-0.55

 Vaugh

-0.55

 rooms

-0.54

POSITIVE LOGITS

 spokesperson

0.58

ozy

0.58

 apology

0.56

 standpoint

0.56

ccording

0.55

 perspective

0.52

 spokeswoman

0.49

 figure

0.49

 Perspective

0.49

 clutch

0.48

Activations Density 1.558%

descriptive and impactful words related to various situations and objects

terms related to negative or challenging situations

No Comments

No Known Activations