INDEX

Explanations

words related to accusations or criticisms

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter "s"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

hyde

-0.79

 Wikileaks

-0.72

 unrestricted

-0.68

 incumbent

-0.64

 WikiLeaks

-0.62

 Sacrament

-0.61

 bases

-0.60

 Scourge

-0.60

 Pharaoh

-0.60

 undecided

-0.59

POSITIVE LOGITS

uss

1.14

outher

1.14

igm

1.13

arah

1.13

agging

1.13

idd

1.13

arin

1.11

iren

1.10

igma

1.10

ams

1.09

Activations Density 0.023%

words related to accusations or criticisms

occurrences of the letter "s"

No Comments

No Known Activations