INDEX

Explanations

terms related to investigation and pressure

oai_token-act-pair · gpt-3.5-turbo

terms associated with investigations, attacks, and surveillance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

atin

-0.72

helle

-0.71

urtle

-0.68

anne

-0.67

apolis

-0.66

pri

-0.65

rael

-0.61

uckland

-0.60

atten

-0.60

ocre

-0.60

POSITIVE LOGITS

ħĭ

1.00

FontSize

0.72

 microscope

0.69

Ú

0.69

Ùİ

0.67

$.

0.67

BAT

0.66

 occasions

0.63

"},"

0.63

izon

0.63

Activations Density 0.065%

terms related to investigation and pressure

terms associated with investigations, attacks, and surveillance

No Comments

No Known Activations