INDEX

Explanations

The neuron is looking for phrases indicating uncertainty or contemplation about a concept or idea

oai_token-act-pair · gpt-3.5-turbo

phrases that express the concept of "what it is."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

teness

-0.84

 Extend

-0.68

oso

-0.67

 Shall

-0.65

ichever

-0.63

 Refresh

-0.61

 Hitch

-0.60

locks

-0.58

uish

-0.57

 pend

-0.57

POSITIVE LOGITS

 supposed

0.84

nt

0.80

olate

0.76

Ĥİ

0.75

 meant

0.75

 actually

0.73

 going

0.73

 akin

0.72

 intrinsically

0.72

 presently

0.72

Activations Density 0.321%

The neuron is looking for phrases indicating uncertainty or contemplation about a concept or idea

phrases that express the concept of "what it is."

No Comments

No Known Activations