INDEX

Explanations

words related to news outlets or media organizations

oai_token-act-pair · gpt-3.5-turbo

references to various media outlets, particularly focusing on 'NET', 'TV', and 'GN'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ovych

-0.78

riks

-0.78

properties

-0.67

 craft

-0.67

ients

-0.66

ient

-0.65

Reviewer

-0.65

settings

-0.65

ply

-0.65

animate

-0.61

POSITIVE LOGITS

HY

0.86

 Insider

0.84

BUS

0.83

azeera

0.80

 NEWS

0.79

ILE

0.74

NEWS

0.73

Ãī

0.73

 DOWN

0.73

CLUS

0.72

Activations Density 0.029%

words related to news outlets or media organizations

references to various media outlets, particularly focusing on 'NET', 'TV', and 'GN'

No Comments

No Known Activations