INDEX

Explanations

phrases related to logs or logging

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of "Log," indicating a focus on logging or documentation topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ovember

-0.81

 panc

-0.74

+++

-0.73

 Blasio

-0.70

hattan

-0.70

zzi

-0.69

cancer

-0.69

uries

-0.68

rael

-0.67

 sunscreen

-0.65

POSITIVE LOGITS

Log

3.84

Log

2.69

log

2.36

LOG

2.29

log

2.06

 logs

1.97

LOG

1.92

 logging

1.67

 logger

1.63

 DevOnline

1.48

Activations Density 0.026%

phrases related to logs or logging

repeated mentions of "Log," indicating a focus on logging or documentation topics

No Comments

No Known Activations