INDEX

Explanations

phrases indicating a detailed analysis or explanation

oai_token-act-pair · gpt-3.5-turbo

terms related to summaries or analyses of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Wee

-0.73

weather

-0.70

bees

-0.66

iatrics

-0.65

gged

-0.63

Na

-0.62

amount

-0.60

Bone

-0.60

ecd

-0.57

Anderson

-0.57

POSITIVE LOGITS

0.99

alore

0.74

aucus

0.73

ership

0.72

sie

0.69

halla

0.69

ilitation

0.69

amily

0.69

stakes

0.68

ed

0.67

Activations Density 0.079%

phrases indicating a detailed analysis or explanation

terms related to summaries or analyses of information

No Comments

No Known Activations