INDEX

Explanations

terms related to institutions or institutionalization, particularly emphasizing negative connotations

oai_token-act-pair · gpt-3.5-turbo

references to institutionalization and systemic issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

kin

-0.75

vous

-0.74

manship

-0.74

bane

-0.74

nen

-0.74

spell

-0.73

llan

-0.72

cius

-0.72

Woman

-0.72

hammad

-0.70

POSITIVE LOGITS

ized

1.36

ised

1.16

ization

1.11

izational

1.08

izable

1.01

izing

0.99

isations

0.99

 inertia

0.98

izations

0.97

 racism

0.97

Activations Density 0.084%

terms related to institutions or institutionalization, particularly emphasizing negative connotations

references to institutionalization and systemic issues

No Comments

No Known Activations