INDEX

Explanations

common nouns referring to different individuals

oai_token-act-pair · gpt-3.5-turbo

references to people involved in a narrative or situation

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

require

-0.68

arning

-0.67

ilver

-0.64

respect

-0.62

olars

-0.62

 Length

-0.62

 Members

-0.61

ourcing

-0.60

ornia

-0.60

ystem

-0.59

POSITIVE LOGITS

 himself

0.87

liest

0.77

stown

0.73

iest

0.72

's

0.71

 claimant

0.69

 regretted

0.67

 herself

0.67

 promptly

0.66

 complainant

0.66

Activations Density 0.285%

common nouns referring to different individuals

references to people involved in a narrative or situation

No Comments

No Known Activations