INDEX

Explanations

names related to a specific political party

oai_token-act-pair · gpt-3.5-turbo

references to political parties and affiliations

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

crow

-0.71

things

-0.70

scene

-0.68

erick

-0.68

stocks

-0.67

 Icar

-0.67

lasses

-0.67

bearing

-0.66

where

-0.66

theless

-0.66

POSITIVE LOGITS

apolis

1.01

ya

0.98

hedral

0.95

endi

0.94

achment

0.94

elson

0.94

Ã±a

0.91

opsis

0.89

qa

0.89

ached

0.86

Activations Density 0.018%

names related to a specific political party

references to political parties and affiliations

No Comments

No Known Activations