INDEX

Explanations

mentions of specific numbers or rankings

oai_token-act-pair · gpt-3.5-turbo

the definite article "the" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

âĢİ

-0.73

leeve

-0.67

Joined

-0.66

 Akin

-0.62

 Fiona

-0.62

fulness

-0.60

èª

-0.60

 Gohan

-0.59

chu

-0.59

imaru

-0.58

POSITIVE LOGITS

 aforementioned

0.97

 latter

0.95

 heaviest

0.92

 same

0.87

 smallest

0.84

 largest

0.81

 proverbial

0.81

 highest

0.79

 utmost

0.79

 worst

0.79

Activations Density 0.360%

mentions of specific numbers or rankings

the definite article "the" in various contexts

No Comments

No Known Activations