INDEX

Explanations

locations of resorts and hotels

oai_token-act-pair · gpt-3.5-turbo

mentions of resorts and related establishments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Ð¼

-0.73

 ignition

-0.68

iple

-0.68

RNA

-0.65

 SIGN

-0.63

 matter

-0.61

Ð²

-0.61

cm

-0.61

matter

-0.60

 filament

-0.60

POSITIVE LOGITS

 Hotel

1.09

 hotels

1.03

 attractions

0.98

pole

0.96

 hotel

0.94

 attraction

0.93

 Resort

0.93

 Plaza

0.89

pav

0.86

 Casino

0.86

Activations Density 0.041%

locations of resorts and hotels

mentions of resorts and related establishments

No Comments

No Known Activations