INDEX

Explanations

references to the term "Ba" with varying activations

oai_token-act-pair · gpt-3.5-turbo

the occurrences of the name "Ba" 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 @ 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

lessly

-0.88

wise

-0.82

otle

-0.77

ments

-0.72

lessness

-0.72

matic

-0.71

mented

-0.68

geist

-0.67

LAND

-0.67

ITAL

-0.66

POSITIVE LOGITS

atar

0.99

uble

0.99

Ba

0.91

uman

0.86

iley

0.84

umann

0.84

ison

0.82

plin

0.81

ild

0.81

iting

0.80

Activations Density 0.012%

references to the term "Ba" with varying activations

the occurrences of the name "Ba" in various contexts

No Comments

No Known Activations