INDEX

Explanations

number patterns represented as sequences of three connected numbers

oai_token-act-pair · gpt-3.5-turbo

numerical data in specific contexts, likely related to scores or statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 jarring

-0.68

 mete

-0.67

 genuine

-0.66

 waning

-0.65

 obvious

-0.65

 consum

-0.65

 mature

-0.64

 treat

-0.64

 glaring

-0.64

 audience

-0.64

POSITIVE LOGITS

1.16

1.15

1.05

1.03

1.00

Switch

0.95

0.93

0.92

quart

0.90

Activations Density 0.065%

number patterns represented as sequences of three connected numbers

numerical data in specific contexts, likely related to scores or statistics

No Comments

No Known Activations