INDEX

Explanations

words related to drug use or references

oai_token-act-pair · gpt-3.5-turbo

terms related to self-inflicted harm or injury

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 dues

-0.74

 fares

-0.69

 grades

-0.69

 Rover

-0.69

ARY

-0.65

 standalone

-0.64

 semester

-0.64

 stag

-0.64

 resumes

-0.64

 Stand

-0.62

POSITIVE LOGITS

inf

4.17

Inf

2.26

Inf

1.65

inf

1.39

inst

1.16

infect

1.14

inc

1.12

inter

1.04

INF

1.01

imm

1.00

Activations Density 0.011%

words related to drug use or references

terms related to self-inflicted harm or injury

No Comments

No Known Activations