INDEX

Explanations

terms related to being a target or victim

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "prey."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

andro

-0.67

PAT

-0.66

leaf

-0.65

arp

-0.65

ANA

-0.62

ATIONAL

-0.61

inness

-0.61

quote

-0.60

essa

-0.60

 Patriot

-0.60

POSITIVE LOGITS

 prey

1.18

 vanquished

0.81

 predators

0.78

odon

0.78

fish

0.77

ãĤ¼ãĤ¦ãĤ¹

0.77

 Horses

0.71

 Whale

0.70

 predator

0.70

acement

0.67

Activations Density 0.006%

terms related to being a target or victim

references to the concept of "prey."

No Comments

No Known Activations