INDEX

Explanations

phrases related to leadership positions and prominent figures

oai_token-act-pair · gpt-3.5-turbo

commas followed by significant numerical values or references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

idian

-0.61

ente

-0.60

hower

-0.59

 Loren

-0.59

olars

-0.57

azel

-0.56

LF

-0.52

Actor

-0.52

olate

-0.51

 Roose

-0.51

POSITIVE LOGITS

into

0.72

 whose

0.72

whose

0.72

 which

0.71

who

0.69

who

0.66

aka

0.64

 whom

0.63

which

0.61

 inviting

0.60

Activations Density 0.288%

phrases related to leadership positions and prominent figures

commas followed by significant numerical values or references

No Comments

No Known Activations