INDEX

Explanations

numbers and numerical patterns such as counts and amounts

oai_token-act-pair · gpt-3.5-turbo

quantitative data and statistics related to performance or metrics

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

tnc

-0.70

 favoured

-0.67

earance

-0.64

erential

-0.64

 protective

-0.64

NetMessage

-0.63

olitical

-0.62

oslav

-0.61

 unwelcome

-0.61

Introduced

-0.59

POSITIVE LOGITS

1.18

 counting

1.15

1.12

1.10

1.08

1.06

Activations Density 0.629%

numbers and numerical patterns such as counts and amounts

quantitative data and statistics related to performance or metrics

No Comments

No Known Activations