INDEX

Explanations

references to political entities or government positions

oai_token-act-pair · gpt-3.5-turbo

references to the United States and its government

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 eleph

-0.73

bet

-0.69

rawdownloadcloneembedreportprint

-0.69

 bending

-0.68

anga

-0.65

+++

-0.59

bda

-0.59

ãĤ¶

-0.59

 supers

-0.59

 circum

-0.58

POSITIVE LOGITS

0.93

0.83

Va

0.83

KS

0.82

Conn

0.81

Ds

0.80

Ku

0.77

State

0.77

orks

0.77

Activations Density 0.060%

references to political entities or government positions

references to the United States and its government

No Comments

No Known Activations