INDEX

Explanations

terms related to tracking or monitoring activities

oai_token-act-pair · gpt-3.5-turbo

references to tracking in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enic

-0.83

 Argument

-0.80

Arg

-0.75

aughs

-0.73

rose

-0.73

ardless

-0.70

 Nare

-0.70

 Flowers

-0.70

 appell

-0.69

 Aven

-0.68

POSITIVE LOGITS

 tracking

3.77

 Tracking

2.89

 tracker

2.54

tracking

2.49

 tracked

2.22

 Tracker

1.82

 monitoring

1.81

 tracing

1.73

 track

1.57

Tracker

1.49

Activations Density 0.016%

terms related to tracking or monitoring activities

references to tracking in various contexts

No Comments

No Known Activations