INDEX

Explanations

terms related to automatic, timed processes or actions

oai_token-act-pair · gpt-3.5-turbo

technical terms and functions related to automated systems and mechanisms

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

wives

-0.85

ylum

-0.83

 ethn

-0.81

blog

-0.80

 blogs

-0.79

 bloggers

-0.77

archives

-0.76

 dating

-0.76

folk

-0.75

Blog

-0.75

POSITIVE LOGITS

 diseng

1.42

 restart

1.26

 braking

1.24

 pause

1.14

 pauses

1.13

 STOP

1.13

 brake

1.07

 momentarily

1.06

 reset

1.05

 brakes

1.05

Activations Density 0.452%

terms related to automatic, timed processes or actions

technical terms and functions related to automated systems and mechanisms

No Comments

No Known Activations