INDEX

Explanations

mentions of specific proper nouns, particularly names and locations such as "Kamp" and "Kemp."

oai_token-act-pair · gpt-3.5-turbo

references to a specific entity or name related to Kamp or Kemp

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

-0.81

EMS

-0.77

*/(

-0.74

ativity

-0.73

.>>

-0.72

 partName

-0.72

 table

-0.70

owler

-0.67

ħĭ

-0.66

çīĪ

-0.64

POSITIVE LOGITS

 Kamp

1.24

aign

0.81

fu

0.77

anga

0.77

lins

0.77

sta

0.75

camp

0.74

 Hels

0.74

oons

0.73

dar

0.72

Activations Density 0.013%

mentions of specific proper nouns, particularly names and locations such as "Kamp" and "Kemp."

references to a specific entity or name related to Kamp or Kemp

No Comments

No Known Activations