INDEX

Explanations

references to locations or places

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of specific locations or places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 pedal

-0.67

 Ratio

-0.61

Arab

-0.61

Rod

-0.61

 decap

-0.60

 indoctr

-0.60

 list

-0.59

 determined

-0.59

 arming

-0.59

 lament

-0.59

POSITIVE LOGITS

oa

3.80

uu

1.80

ua

1.33

owa

1.21

oji

1.08

 Gaga

1.00

ui

0.97

aho

0.95

oj

0.94

anta

0.94

Activations Density 0.007%

references to locations or places

repeated mentions of specific locations or places

No Comments

No Known Activations