INDEX

Explanations

the phrase "hard to" followed by a verb

oai_token-act-pair · gpt-3.5-turbo

phrases that repeatedly include the word "to" indicating intentions or actions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

afety

-0.93

minus

-0.68

raped

-0.67

 Neuroscience

-0.64

written

-0.63

 Culture

-0.62

plates

-0.61

iquette

-0.60

cand

-0.60

ements

-0.60

POSITIVE LOGITS

 achieve

1.13

 reconcile

1.07

 imagine

1.06

 accomplish

1.06

 distinguish

1.05

 obtain

1.05

 attain

1.05

 penetrate

1.04

 reach

1.01

 identify

0.98

Activations Density 0.100%

the phrase "hard to" followed by a verb

phrases that repeatedly include the word "to" indicating intentions or actions

No Comments

No Known Activations