INDEX

Explanations

numerical counts and quantifiers

oai_token-act-pair · gpt-3.5-turbo

references to counting and quantifying actions or items

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

avage

-0.82

hement

-0.75

ĨĴ

-0.73

forum

-0.72

\\\\\\\\

-0.70

nered

-0.70

reen

-0.67

aline

-0.66

redo

-0.62

ãĥ³ãĤ¸

-0.62

POSITIVE LOGITS

 elapsed

0.82

 probabilities

0.78

 positives

0.72

 milliseconds

0.71

 blessings

0.70

 digits

0.69

abytes

0.68

Quantity

0.68

 hairs

0.67

 beads

0.67

Activations Density 0.152%

numerical counts and quantifiers

references to counting and quantifying actions or items

No Comments

No Known Activations