INDEX

Explanations

information related to legal or political matters

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

segments of text that indicate significant events or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

ses

-0.81

lessly

-0.65

 elusive

-0.64

 choosing

-0.62

ovich

-0.62

 periodic

-0.61

 manif

-0.60

 portfolios

-0.60

owsky

-0.59

 spoiled

-0.58

POSITIVE LOGITS

Ct

0.90

IPS

0.85

RNA

0.84

 ILCS

0.83

DAY

0.83

NYSE

0.80

UTION

0.80

Ire

0.79

Activations Density 0.019%

information related to legal or political matters

segments of text that indicate significant events or actions

No Comments

No Known Activations