INDEX

Explanations

proper nouns related to various entities, particularly the name "Freeman" at different levels of activation

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals, particularly someone named Freeman

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ificial

-0.82

ģĸ

-0.77

aron

-0.75

alid

-0.74

rolog

-0.70

assic

-0.69

oiler

-0.68

ãĤ°

-0.67

benefit

-0.66

 Mehran

-0.64

POSITIVE LOGITS

 Freeman

0.80

ufact

0.73

lings

0.71

vernment

0.70

iquette

0.69

ason

0.68

burg

0.67

 Duff

0.67

boro

0.66

 Sketch

0.66

Activations Density 0.012%

proper nouns related to various entities, particularly the name "Freeman" at different levels of activation

mentions of specific individuals, particularly someone named Freeman

No Comments

No Known Activations