INDEX

Explanations

numerical sequences aligned to the number four

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "four" and related numerical references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ultan

-0.67

INS

-0.65

OCK

-0.64

merga

-0.63

Dri

-0.62

 Notes

-0.62

MO

-0.59

ushed

-0.58

idges

-0.57

 Flight

-0.57

POSITIVE LOGITS

theless

1.13

lihood

1.08

etheless

0.89

hesda

0.85

hyde

0.83

tin

0.81

tesy

0.79

hole

0.78

ieth

0.78

four

0.77

Activations Density 0.016%

numerical sequences aligned to the number four

occurrences of the word "four" and related numerical references

No Comments

No Known Activations