INDEX

Explanations

words related to investigation or deep exploration

oai_token-act-pair · gpt-3.5-turbo

words related to exploration or delving into topics

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

 Duff

-0.63

 ounce

-0.63

ãĥ¼ãĥĨãĤ£

-0.62

pak

-0.62

atsu

-0.61

 lining

-0.61

loading

-0.60

nance

-0.59

agen

-0.59

asu

-0.59

POSITIVE LOGITS

schild

0.76

ocobo

0.74

 Mour

0.73

ython

0.67

 Parables

0.66

letters

0.65

osate

0.63

bre

0.63

resy

0.63

izons

0.62

Activations Density 0.000%

words related to investigation or deep exploration

words related to exploration or delving into topics

No Comments

No Known Activations