INDEX

Explanations

names of individuals, likely related to law enforcement or legal contexts

oai_token-act-pair · gpt-3.5-turbo

proper names, particularly those of people involved in legal or law enforcement contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Sahara

-0.77

 Boko

-0.77

acters

-0.76

 Antarctica

-0.73

HMS

-0.70

 Georgian

-0.70

ãģ¯

-0.69

tumblr

-0.69

 rebounds

-0.69

 Crimean

-0.68

POSITIVE LOGITS

 oversaw

0.95

shall

0.94

 overseeing

0.90

 oversees

0.89

zynski

0.85

uzz

0.83

meier

0.83

oval

0.83

nel

0.82

oyer

0.82

Activations Density 0.269%

names of individuals, likely related to law enforcement or legal contexts

proper names, particularly those of people involved in legal or law enforcement contexts

No Comments

No Known Activations