INDEX

Explanations

appropriate nouns and proper nouns

oai_token-act-pair · gpt-3.5-turbo

phrases that refer to significant people, events, or entities in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ruction

-0.61

aughter

-0.57

 Coverage

-0.56

"],

-0.55

[];

-0.55

ombat

-0.54

EStreamFrame

-0.54

lly

-0.54

rations

-0.53

POSITIVE LOGITS

aka

0.88

 albeit

0.87

 alas

0.80

 hitherto

0.72

 unlike

0.72

 namely

0.70

 fearing

0.70

 coupled

0.69

whatever

0.67

 despite

0.67

Activations Density 0.295%

appropriate nouns and proper nouns

phrases that refer to significant people, events, or entities in context

No Comments

No Known Activations