INDEX

Explanations

phrases related to decision-making or intention, such as desires, intends, and avoids

oai_token-act-pair · gpt-3.5-turbo

phrases that express desires or intentions

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

é¾įå

-0.71

cffff

-0.70

acia

-0.69

ntil

-0.64

ãĤĮ

-0.62

yd

-0.61

orget

-0.60

Ļ

-0.58

anwhile

-0.58

çīĪ

-0.57

POSITIVE LOGITS

 itself

1.25

its

0.88

ITS

0.76

Its

0.75

Its

0.73

ibly

0.68

 churn

0.64

 leaps

0.64

rive

0.62

ublished

0.61

Activations Density 0.785%

phrases related to decision-making or intention, such as desires, intends, and avoids

phrases that express desires or intentions

No Comments

No Known Activations