INDEX

Explanations

TV news channels and programs

oai_token-act-pair · gpt-3.5-turbo

references to television networks and news outlets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

itance

-0.74

ktop

-0.73

uten

-0.68

awa

-0.68

sell

-0.67

urion

-0.66

abouts

-0.66

notations

-0.64

ebook

-0.64

 hiber

-0.64

POSITIVE LOGITS

ILLE

1.22

NZ

1.00

 Channel

0.91

News

0.88

Channel

0.87

 News

0.83

 NEWS

0.82

 channel

0.81

LV

0.77

WT

0.75

Activations Density 0.038%

TV news channels and programs

references to television networks and news outlets

No Comments

No Known Activations