INDEX

Explanations

words related to governmental or organizational entities or positions

oai_token-act-pair · gpt-3.5-turbo

letters and single characters that may indicate formatting or coding aspects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

å§«

-0.80

schild

-0.66

 Metatron

-0.65

IZE

-0.65

umenthal

-0.64

 sidx

-0.61

LEASE

-0.60

Guest

-0.59

 bell

-0.59

 cheers

-0.58

POSITIVE LOGITS

orio

0.89

ule

0.85

uta

0.83

ara

0.80

anish

0.77

any

0.77

ula

0.74

oda

0.74

ira

0.73

yssey

0.73

Activations Density 0.133%

words related to governmental or organizational entities or positions

letters and single characters that may indicate formatting or coding aspects

No Comments

No Known Activations