INDEX

Explanations

mentions of the name "ima" in various contexts, potentially related to a specific person or place

oai_token-act-pair · gpt-3.5-turbo

instances of names or titles related to individuals and cultural references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

some

-0.87

ridges

-0.82

bread

-0.81

nell

-0.81

rooms

-0.79

pocket

-0.79

draw

-0.76

lies

-0.76

olds

-0.75

houses

-0.74

POSITIVE LOGITS

 Centauri

0.97

ption

0.86

Ã±

0.84

estro

0.80

imi

0.72

Ã±a

0.70

edia

0.70

pora

0.70

uthor

0.70

escal

0.70

Activations Density 0.013%

mentions of the name "ima" in various contexts, potentially related to a specific person or place

instances of names or titles related to individuals and cultural references

No Comments

No Known Activations