INDEX

Explanations

expressions related to coherence

oai_token-act-pair · gpt-3.5-turbo

terms related to coherence and cohesion in ideas or narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 awarded

-0.72

yer

-0.70

abbit

-0.67

enge

-0.67

 derby

-0.62

gio

-0.61

AUT

-0.59

marks

-0.59

 Rober

-0.58

fish

-0.58

POSITIVE LOGITS

ly

1.17

ively

1.12

 coherent

1.06

herent

1.04

herence

0.88

nesses

0.87

 inco

0.85

mble

0.79

 cohesion

0.78

 nodd

0.77

Activations Density 0.015%

expressions related to coherence

terms related to coherence and cohesion in ideas or narratives

No Comments

No Known Activations