INDEX

Explanations

words related to political bills and activities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

phrases indicating legislative processes or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

=-=-=-=-

-0.88

paralle

-0.79

sonian

-0.77

ĸļ

-0.74

gencies

-0.73

role

-0.72

capacity

-0.71

ventures

-0.71

venture

-0.71

riott

-0.70

POSITIVE LOGITS

 debated

1.32

 vetoed

1.27

 defeated

1.13

 enacted

1.12

 stalled

1.11

 ratified

1.10

 passed

1.08

 filib

1.07

 signed

1.05

 voted

1.04

Activations Density 0.144%

words related to political bills and activities

phrases indicating legislative processes or actions

No Comments

No Known Activations