INDEX

Explanations

instances of the word "plan" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to action plans or strategies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

"]=>

-0.87

¿

-0.70

Reading

-0.69

ľ

-0.69

igious

-0.68

ences

-0.68

lez

-0.67

entious

-0.67

laughter

-0.67

inguished

-0.66

POSITIVE LOGITS

 plan

3.70

 plans

2.60

 Plan

2.59

plan

2.46

 Plans

2.34

Plan

2.25

 PLAN

2.14

 planned

1.67

 intend

1.66

 proposal

1.65

Activations Density 0.022%

instances of the word "plan" in various contexts

references to action plans or strategies

No Comments

No Known Activations