INDEX

Explanations

specific names, especially surnames, with a higher preference for the surname "Schmidt."

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals, particularly those with the last name "Schmidt" or "Weber."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ership

-0.98

drawn

-0.74

rity

-0.73

busters

-0.72

lv

-0.72

icative

-0.71

itives

-0.71

draw

-0.70

eas

-0.69

Indian

-0.69

POSITIVE LOGITS

 Schmidt

1.09

acher

0.94

midt

0.87

 Christensen

0.79

meyer

0.76

kson

0.74

bone

0.72

 Herz

0.71

 Schwarz

0.71

bach

0.70

Activations Density 0.017%

specific names, especially surnames, with a higher preference for the surname "Schmidt."

mentions of specific individuals, particularly those with the last name "Schmidt" or "Weber."

No Comments

No Known Activations