INDEX

Explanations

words related to proper nouns or personal names, especially those with emphasis or focus

oai_token-act-pair · gpt-3.5-turbo

specific letters or syllables that appear frequently in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 fluor

-0.69

ailability

-0.64

¿½

-0.62

 grate

-0.61

facing

-0.61

ãĤ´ãĥ³

-0.61

 commute

-0.58

dfx

-0.58

dden

-0.57

BUR

-0.56

POSITIVE LOGITS

uay

0.98

vernment

0.87

andise

0.85

orno

0.83

eta

0.79

ultural

0.77

iru

0.74

nostic

0.71

udo

0.70

zx

0.70

Activations Density 0.293%

words related to proper nouns or personal names, especially those with emphasis or focus

specific letters or syllables that appear frequently in various contexts

No Comments

No Known Activations