INDEX

Explanations

instances where a character takes an action that leads to a specific outcome

oai_token-act-pair · gpt-3.5-turbo

the repeated phrase "only" followed by varying negative or adverse circumstances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

cv

-0.64

iller

-0.62

insula

-0.61

ahead

-0.61

rect

-0.60

ahime

-0.59

lass

-0.59

ji

-0.58

CV

-0.58

atical

-0.57

POSITIVE LOGITS

 kidding

0.93

 marginally

0.89

 minutes

0.79

 weeks

0.76

 seconds

0.76

 incidentally

0.75

 spor

0.73

 realizing

0.73

 moments

0.73

 intermitt

0.72

Activations Density 0.049%

instances where a character takes an action that leads to a specific outcome

the repeated phrase "only" followed by varying negative or adverse circumstances

No Comments

No Known Activations