INDEX

Explanations

mentions of social media posts specifically of the action of tweeting

oai_token-act-pair · gpt-3.5-turbo

instances of the word "tweeted" and its variations, indicating social media activity

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

 comprom

-0.79

cised

-0.69

inventoryQuantity

-0.67

vantage

-0.64

cius

-0.61

glomer

-0.61

ãĥĺ

-0.61

vernment

-0.61

ãĥ´

-0.60

ricular

-0.60

POSITIVE LOGITS

"@

1.12

 condolences

0.98

"#

0.94

 congratulations

0.94

 encouragement

0.85

 pics

0.84

 sarcast

0.83

 angrily

0.83

 screenshots

0.83

 photos

0.82

Activations Density 0.053%

mentions of social media posts specifically of the action of tweeting

instances of the word "tweeted" and its variations, indicating social media activity

No Comments

No Known Activations