INDEX

Explanations

quantitative phrases or measures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.8.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oris

-0.72

roit

-0.68

SEE

-0.67

successfully

-0.63

NET

-0.62

rises

-0.62

 indu

-0.62

arta

-0.60

spring

-0.60

awn

-0.60

POSITIVE LOGITS

 rubble

1.02

 obscurity

0.99

 ashes

0.99

 mere

0.92

 scraps

0.89

 negligible

0.89

 zero

0.86

 insignificant

0.86

 fractions

0.85

 fragments

0.83

Activations Density 0.445%

quantitative phrases or measures

No Comments

No Known Activations