INDEX

Explanations

words related to revisions or updates

oai_token-act-pair · gpt-3.5-turbo

instances of the word "revised" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

acht

-0.76

Craft

-0.69

OPA

-0.64

 charism

-0.64

bara

-0.64

chu

-0.63

ech

-0.63

 Winged

-0.63

asts

-0.61

ecause

-0.61

POSITIVE LOGITS

 guiActiveUn

0.80

 revision

0.78

itized

0.75

 revisions

0.75

 formulations

0.71

 wording

0.68

 retrospect

0.67

 versions

0.67

 equilibrium

0.65

atively

0.64

Activations Density 0.016%

words related to revisions or updates

instances of the word "revised" and its variations

No Comments

No Known Activations