INDEX

Explanations

proper nouns related to people's names

oai_token-act-pair · gpt-3.5-turbo

occurrences of names and terms related to characters or specific themes in a narrative context

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

ngth

-0.83

£ı

-0.75

 showc

-0.74

 kindred

-0.73

¬¼

-0.67

 medial

-0.65

 disparate

-0.65

ĸļ

-0.64

effic

-0.64

 tumult

-0.64

POSITIVE LOGITS

swick

0.77

ucky

0.69

orst

0.68

lang

0.68

atch

0.67

itte

0.67

etz

0.67

cycle

0.65

skin

0.64

chain

0.63

Activations Density 0.215%

proper nouns related to people's names

occurrences of names and terms related to characters or specific themes in a narrative context

No Comments

No Known Activations