INDEX

Explanations

references to achieving success or winning in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "victory" and related contexts

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

 pores

-0.74

asus

-0.71

 notor

-0.69

ridges

-0.68

thening

-0.65

 suscept

-0.62

Distance

-0.62

missible

-0.61

zie

-0.60

 borrowing

-0.60

POSITIVE LOGITS

antly

0.86

hower

0.76

cry

0.75

bringer

0.74

stroke

0.71

 Lann

0.69

 victory

0.68

ceed

0.68

 defeat

0.68

abeth

0.67

Activations Density 0.037%

references to achieving success or winning in various contexts

instances of the word "victory" and related contexts

No Comments

No Known Activations