INDEX

Explanations

specific strings related to episodes of a podcast or TV show

oai_token-act-pair · gpt-3.5-turbo

references to specific episodes of shows or podcasts

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

 bush

-0.71

 shells

-0.70

 marsh

-0.68

 conc

-0.67

 bank

-0.67

 firewall

-0.67

 gard

-0.66

 beaut

-0.66

 deter

-0.66

 range

-0.64

POSITIVE LOGITS

 Episode

3.84

Episode

2.63

episode

1.82

 episode

1.60

isodes

1.47

 episodes

1.45

 Podcast

1.43

Ep

1.36

 Chapter

1.36

isode

1.25

Activations Density 0.010%

specific strings related to episodes of a podcast or TV show

references to specific episodes of shows or podcasts

No Comments

No Known Activations