INDEX

Explanations

names of specific police departments

oai_token-act-pair · gpt-3.5-turbo

references to police departments and their associated officers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

sed

-0.90

actionDate

-0.79

bent

-0.78

isin

-0.78

drive

-0.78

ãĥ¼ãĤ¯

-0.77

piring

-0.75

worldly

-0.74

ulative

-0.74

hend

-0.73

POSITIVE LOGITS

 Officer

0.88

 officer

0.87

 Officers

0.85

 Armory

0.84

 officers

0.82

 Commissioner

0.81

 commissioner

0.79

 Superintendent

0.79

 brass

0.78

 superintendent

0.77

Activations Density 0.076%

names of specific police departments

references to police departments and their associated officers

No Comments

No Known Activations