INDEX

Explanations

names of medical publications or technical terms

oai_token-act-pair · gpt-3.5-turbo

variations of the word "exploit" and related forms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yip

-0.89

channelAvailability

-0.68

afety

-0.68

ritic

-0.67

dim

-0.67

finding

-0.66

pring

-0.65

senal

-0.64

ensional

-0.63

 insecure

-0.61

POSITIVE LOGITS

rocal

0.85

 Tayyip

0.83

iencies

0.81

atorium

0.81

ators

0.80

itative

0.74

issance

0.74

aughters

0.73

cled

0.72

itely

0.72

Activations Density 0.107%

names of medical publications or technical terms

variations of the word "exploit" and related forms

No Comments

No Known Activations