INDEX

Explanations

the name "Uma" in various contexts and mentions

oai_token-act-pair · gpt-3.5-turbo

the recurrence of the term "Zuma" related to various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 check

-0.77

 checks

-0.64

ply

-0.61

 invent

-0.61

 discoveries

-0.61

 totals

-0.60

 stocking

-0.60

 Beck

-0.60

eas

-0.60

 pred

-0.59

POSITIVE LOGITS

uma

4.68

umatic

1.99

umat

1.69

um

1.20

uman

1.18

umi

1.18

umo

1.16

ema

1.14

ugi

1.14

uro

1.13

Activations Density 0.011%

the name "Uma" in various contexts and mentions

the recurrence of the term "Zuma" related to various contexts

No Comments

No Known Activations