INDEX

Explanations

quotations from individuals

oai_token-act-pair · gpt-3.5-turbo

utilization of quotations in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ãĥ¼ãĥĨ

-0.71

venth

-0.69

ãĥ¼ãĥĨãĤ£

-0.64

ems

-0.59

ensible

-0.58

acly

-0.57

filled

-0.57

irrel

-0.56

 conflic

-0.55

Honest

-0.55

POSITIVE LOGITS

0.75

:"

0.74

 goodbye

0.73

:]

0.71

:'

0.71

 "...

0.68

jriwal

0.67

 Rohing

0.66

 sarcast

0.65

]:

0.65

Activations Density 0.189%

quotations from individuals

utilization of quotations in the text

No Comments

No Known Activations