INDEX

Explanations

instances where something is being perceived or interpreted by others

oai_token-act-pair · gpt-3.5-turbo

instances of the word "viewed" in various contexts and formats

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

 promise

-0.67

prov

-0.62

 overtime

-0.61

HK

-0.60

bass

-0.60

rh

-0.57

 Promise

-0.57

plex

-0.57

 Noise

-0.56

hh

-0.56

POSITIVE LOGITS

 viewed

3.54

 regarded

1.88

 seen

1.58

 accessed

1.57

 view

1.51

 viewing

1.49

 construed

1.49

 interpreted

1.48

 watched

1.48

view

1.46

Activations Density 0.010%

instances where something is being perceived or interpreted by others

instances of the word "viewed" in various contexts and formats

No Comments

No Known Activations