INDEX

Explanations

phrases related to political entities, specifically the "House of Representatives."

oai_token-act-pair · gpt-3.5-turbo

references to the House of Representatives

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

ãĤ¨ãĥ«

-1.00

istically

-0.84

zsche

-0.74

istic

-0.74

fuel

-0.72

capital

-0.72

illas

-0.71

clad

-0.68

cash

-0.68

company

-0.66

POSITIVE LOGITS

 Representatives

1.00

 Speaker

0.94

 Appropriations

0.88

OTUS

0.79

hips

0.78

 Office

0.78

woman

0.77

ricting

0.75

 Representative

0.74

 Commissioners

0.74

Activations Density 0.020%

phrases related to political entities, specifically the "House of Representatives."

references to the House of Representatives

No Comments

No Known Activations