INDEX

Explanations

proper nouns related to individuals and political events

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to a specific individual or title associated with leadership or authority

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 parting

-0.74

 diver

-0.65

 linem

-0.64

 sidx

-0.64

 Scouting

-0.61

 contrasting

-0.61

DERR

-0.61

ãĥ¼ãĥĨ

-0.60

 spacing

-0.60

 prominently

-0.59

POSITIVE LOGITS

otte

1.16

meric

1.11

ahu

1.10

alon

1.10

ubi

1.09

reon

1.04

atana

1.04

cock

1.01

cliffe

0.97

ubs

0.94

Activations Density 0.009%

proper nouns related to individuals and political events

references to a specific individual or title associated with leadership or authority

No Comments

No Known Activations