INDEX

Explanations

instances of collaborative activities or projects

oai_token-act-pair · gpt-3.5-turbo

terms related to collaboration

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

leigh

-0.82

 Sabha

-0.78

 Adin

-0.73

¥µ

-0.68

politics

-0.68

den

-0.67

gm

-0.67

va

-0.66

seed

-0.66

corruption

-0.65

POSITIVE LOGITS

 collaborations

1.00

 collaborators

0.97

 collaboration

0.93

 collaborator

0.90

 collabor

0.87

 arrang

0.87

 collaborating

0.84

 collaborative

0.83

 collaborate

0.83

 partner

0.82

Activations Density 0.017%

instances of collaborative activities or projects

terms related to collaboration

No Comments

No Known Activations