INDEX

Explanations

negative statements or denials containing the word "not"

oai_token-act-pair · gpt-3.5-turbo

negations or phrases indicating something is not acceptable or favorable

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

çīĪ

-0.84

çļ

-0.77

éĹĺ

-0.72

ãĤ©

-0.71

umbnail

-0.68

States

-0.68

 Cosponsors

-0.67

ennes

-0.66

former

-0.65

¿½

-0.65

POSITIVE LOGITS

 advisable

1.20

 conducive

1.18

 acceptable

1.18

 necessarily

1.17

 feasible

1.17

 happening

1.13

hin

1.12

 gonna

1.11

 uncommon

1.09

 enough

1.06

Activations Density 0.103%

negative statements or denials containing the word "not"

negations or phrases indicating something is not acceptable or favorable

No Comments

No Known Activations