INDEX

Explanations

negative or critical language

oai_token-act-pair · gpt-3.5-turbo

phrases related to strong emotions or distressing situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

DragonMagazine

-0.62

Moz

-0.56

 Squadron

-0.52

Org

-0.51

 Hound

-0.51

OLD

-0.51

Hin

-0.50

KH

-0.50

 Bulk

-0.49

 Enhancement

-0.49

POSITIVE LOGITS

$.

0.87

'.

0.81

`.

0.77

_.

0.76

''.

0.75

%.

0.71

*.

0.69

.''

0.67

().

0.67

 detriment

0.66

Activations Density 1.526%

negative or critical language

phrases related to strong emotions or distressing situations

No Comments

No Known Activations