INDEX

Explanations

references to radio stations

oai_token-act-pair · gpt-3.5-turbo

mentions of radio, particularly specific radio shows and stations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

rg

-0.79

-0.76

resses

-0.70

gets

-0.70

URES

-0.67

aunder

-0.66

jamin

-0.65

erest

-0.62

ean

-0.62

aws

-0.62

POSITIVE LOGITS

active

1.06

 Radio

1.02

Radio

0.98

actively

0.93

ode

0.91

 Shack

0.91

 antennas

0.90

radio

0.85

 stations

0.84

 iodine

0.83

Activations Density 0.006%

references to radio stations

mentions of radio, particularly specific radio shows and stations

No Comments

No Known Activations