INDEX

Explanations

The main thing this neuron does is find mentions of specific dates related to news events or announcements

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and conjunctions in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

↵Âł

-0.57

.*

-0.55

 doesnt

-0.49

ie

-0.48

_.

-0.48

.(

-0.47

np

-0.47

*.

-0.46

**

-0.46

(*

-0.46

POSITIVE LOGITS

 SpaceX

0.49

 NASCAR

0.49

 Canaver

0.49

 Charlottesville

0.48

 wildfires

0.46

 Ivanka

0.46

 Seahawks

0.46

 astronomers

0.45

 Hillary

0.44

 Melania

0.44

Activations Density 1.945%

The main thing this neuron does is find mentions of specific dates related to news events or announcements

punctuation marks and conjunctions in the text

No Comments

No Known Activations