INDEX

Explanations

names or terms related to "Haynes."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "nes"

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

ivation

-0.72

arna

-0.67

atories

-0.66

amiya

-0.64

irlf

-0.64

 prefrontal

-0.62

top

-0.62

 anecdotal

-0.62

 hers

-0.62

masters

-0.61

POSITIVE LOGITS

earch

1.06

nes

0.97

ville

0.95

boro

0.93

cape

0.88

itters

0.88

ade

0.87

erve

0.87

creen

0.87

esis

0.87

Activations Density 0.009%

names or terms related to "Haynes."

occurrences of the substring "nes"

No Comments

No Known Activations