INDEX

Explanations

expressions related to potential outcomes or consequences

oai_token-act-pair · gpt-3.5-turbo

phrases that discuss potential future events and outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

cius

-0.62

è¦ļéĨĴ

-0.60

edia

-0.57

 analogy

-0.57

 Lear

-0.57

 Shuttle

-0.57

Sher

-0.56

DX

-0.55

plom

-0.55

 reminder

-0.53

POSITIVE LOGITS

enance

0.93

 someday

0.90

 anytime

0.79

lead

0.78

 unless

0.76

uate

0.76

if

0.71

 THEM

0.68

 tomorrow

0.68

ect

0.66

Activations Density 0.428%

expressions related to potential outcomes or consequences

phrases that discuss potential future events and outcomes

No Comments

No Known Activations